Imputation and estimation under nonignorable nonresponse for
household surveys with missing covariate information
Imputation and estimation under nonignorable nonresponse for
household surveys with missing covariate information
In this paper we develop and apply new methods for handling not missing at random
(NMAR) nonresponse. We assume a model for the outcome variable under complete response
and a model for the response probability, which is allowed to depend on the outcome and
auxiliary variables. The two models define the model holding for the outcomes observed for the
responding units, which can be tested. Our methods utilize information on the population totals
of some or all of the auxiliary variables in the two models, but we do not require that the
auxiliary variables are observed for the nonresponding units. We develop an algorithm for
estimating the parameters governing the two models and show how to estimate the distributions
of the missing covariates and outcomes, which are then used for imputing the missing values for
the nonresponding units and for estimating population means and the variances of the estimators.
We also consider several test statistics for testing the model fitted to the observed data and study
their performance, thus validating the proposed procedure. The new developments are illustrated
using simulated data and a real data set collected as part of the Household Expenditure Survey
carried out by the Israel Central Bureau of Statistics in 2005.
bootstrap, calibration, Horvitz-Thompson type estimator, nonrespondents distribution, respondents distribution
Southampton Statistical Sciences Research Institute, University of Southampton
Pfeffermann, Danny
c7fe07a0-9715-42ce-b90b-1d4f2c2c6ffc
Sikov, Anna
81a74f0d-d006-49df-80f5-ed626b989828
21 June 2010
Pfeffermann, Danny
c7fe07a0-9715-42ce-b90b-1d4f2c2c6ffc
Sikov, Anna
81a74f0d-d006-49df-80f5-ed626b989828
Pfeffermann, Danny and Sikov, Anna
(2010)
Imputation and estimation under nonignorable nonresponse for
household surveys with missing covariate information
(S3RI Methodology Working Papers, M10/04)
Southampton, GB.
Southampton Statistical Sciences Research Institute, University of Southampton
Record type:
Monograph
(Working Paper)
Abstract
In this paper we develop and apply new methods for handling not missing at random
(NMAR) nonresponse. We assume a model for the outcome variable under complete response
and a model for the response probability, which is allowed to depend on the outcome and
auxiliary variables. The two models define the model holding for the outcomes observed for the
responding units, which can be tested. Our methods utilize information on the population totals
of some or all of the auxiliary variables in the two models, but we do not require that the
auxiliary variables are observed for the nonresponding units. We develop an algorithm for
estimating the parameters governing the two models and show how to estimate the distributions
of the missing covariates and outcomes, which are then used for imputing the missing values for
the nonresponding units and for estimating population means and the variances of the estimators.
We also consider several test statistics for testing the model fitted to the observed data and study
their performance, thus validating the proposed procedure. The new developments are illustrated
using simulated data and a real data set collected as part of the Household Expenditure Survey
carried out by the Israel Central Bureau of Statistics in 2005.
Text
s3ri-workingpaper-M10-04.pdf
- Other
More information
Published date: 21 June 2010
Keywords:
bootstrap, calibration, Horvitz-Thompson type estimator, nonrespondents distribution, respondents distribution
Identifiers
Local EPrints ID: 158453
URI: http://eprints.soton.ac.uk/id/eprint/158453
PURE UUID: 674e4e5b-7587-4e60-87a3-d229a6076534
Catalogue record
Date deposited: 23 Jun 2010 08:30
Last modified: 14 Mar 2024 01:50
Export record
Contributors
Author:
Anna Sikov
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics